Gemini SEO: What Google Documents About Ranking in Gemini

Gemini SEO is the work of getting your pages used as sources by Google's Gemini models — in AI Overviews, in AI Mode, and in the Gemini app. Those three surfaces are not one thing, and that is the whole difficulty. Two of them are Search features: Google documents that a page is eligible to appear in them if it is indexed and eligible to be shown with a snippet, and states there are “no additional technical requirements.” The third, the Gemini app, is governed by a separate robots.txt token, Google-Extended. Google documents that this token “does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search.” So the same robots.txt line that removes you from one surface leaves the other two untouched. Everything sold under this heading reduces to knowing which surface you are talking about.

What Gemini SEO actually means

Gemini is a family of models, not a search engine. It sits underneath several products that behave differently, and the phrase collapses them together. When someone asks about Gemini SEO they usually mean one of three jobs: appearing as a supporting link in an AI Overview, appearing in AI Mode, or being retrieved when someone asks the Gemini app a question. The first two are Google Search features. The third is a separate product with its own opt-out.

This distinction is not pedantry — it decides which control you reach for. Google states plainly that “AI is built into Search and integral to how Search functions, which is why robots.txt directives for Googlebot is the control for site owners to manage access to how their sites are crawled for Search.” (Documented — Google, “AI features and your website”.) There is no Gemini crawler you can allow to buy your way into AI Overviews, because AI Overviews read the same index Googlebot builds.

A note on the older name. Google renamed Bard to Gemini in February 2024, and a good deal of advice still circulating under Bard SEO was written for that earlier product. The rename did not change the mechanism — the app grounded its answers in Google's index then, and it does now — but it did change the vocabulary, and any Bard SEO checklist that told you to allow a dedicated “Bard crawler” was describing something that never existed. The control token was Google-Extended before the rename and it still is.

Three surfaces, three sets of rules

Surface	What decides whether you appear	What removes you
AI Overviews (in Search)	Indexed by Googlebot and eligible to be shown with a snippet. Google documents no additional technical requirements.	noindex, nosnippet, data-nosnippet, max-snippet, or disallowing Googlebot.
AI Mode (in Search)	The same eligibility rule. Google notes AI Mode and AI Overviews “may use different models and techniques,” so the links shown vary.	The same snippet and indexing controls.
Gemini app (gemini.google.com)	Grounding — Google's word for supplying content from the Search index to the model at prompt time.	Google-Extended in robots.txt.

One more token is worth naming so you do not block it by accident. Google-CloudVertexBot crawls sites at a site owner's own request for building Vertex AI Agents; Google documents that it “has no effect on Google Search or other products.” It is not part of Gemini SEO in any practical sense, and it is not a route into AI Overviews.

What Google-Extended controls, and what it does not

Google-Extended is unusual: it is a control token rather than a crawler. Google documents that it “doesn't have a separate HTTP request user agent string,” and that “crawling is done with existing Google user agent strings; the robots.txt user-agent token is used in a control capacity.” You will therefore never see it in your server logs, which is why so many audits report it as inactive.

What it governs, in Google's own words, is “whether content Google crawls from their sites may be used for training future generations of Gemini models that power Gemini Apps and Vertex AI API for Gemini and for grounding (providing content from the Google Search index to the model at prompt time to improve factuality and relevancy) in Gemini Apps.” (Documented — Google, “Google's common crawlers”.)

Read that carefully, because it carries both halves of the decision. Blocking Google-Extended does cost you something real: grounding in the Gemini app, which is the mechanism by which the app cites a live page rather than answering from memory. And it costs you nothing in Search, because Google states the token “does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search.” Most advice gets exactly one of those two halves right.

Documented, versus widely believed

Four claims circulate constantly. Google's documentation contradicts all four.

Widely believed	What Google documents
Blocking Google-Extended removes you from AI Overviews.	It “does not impact a site's inclusion in Google Search nor is it used as a ranking signal.” AI Overviews are governed by Googlebot and the snippet controls.
Blocking Google-Extended costs nothing, so block it by default.	It governs grounding in Gemini Apps. Blocking it removes the app's ability to ground answers in your pages.
You need an AI-specific file or schema to be cited.	“You don't need to create new machine readable files, AI text files, or markup… There's also no special schema.org structured data that you need to add.”
There is a separate optimisation discipline for AI Overviews.	“There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.”

That last one deserves care rather than celebration. “No additional requirements” means the entry ticket is ordinary indexing — it does not mean the outcome is ordinary. Which of the eligible pages actually gets used is a retrieval and synthesis decision Google does not document, and nobody outside Google can report on it honestly. Anyone selling you a Gemini ranking factor is selling you an inference labelled as a fact.

How to rank in Gemini

Because eligibility is ordinary, the work is ordinary — done properly. Four things carry disproportionate weight.

Keep the snippet controls off the pages you want cited. Eligibility requires that a page can be shown with a snippet. A blanket nosnippet or an aggressive max-snippet applied years ago for a different reason will quietly remove a page from every AI surface in Search. Check this before anything else — see indexation and crawlability.

Write passages that survive being lifted. A generative answer quotes a span, not a page. A paragraph that answers its own heading without depending on the three paragraphs above it is the unit that gets used. This is the same craft that generative engine optimization describes engine-agnostically.

Cover the sub-questions, not just the query. Google documents that both AI features may use a “query fan-out” technique — “issuing multiple related searches across subtopics and data sources” — and that this lets them show a wider set of links than classic Search. A page that resolves several adjacent sub-questions has more surfaces to be retrieved on than a page that answers one.

Decide the Google-Extended question deliberately. It is a policy choice about training and app grounding, not an SEO lever. Make it once, write down why, and stop revisiting it in ranking reviews.

What you can measure, and what you cannot

You can measure less than the tools imply. Google documents that sites appearing in AI features are “included in the overall search traffic in Search Console,” reported inside the Performance report under the “Web” search type — which means AI Overview and AI Mode impressions are mixed into the same numbers as ordinary blue links, with no breakdown that separates them. There is no AI Overviews report to open.

What that leaves is inference from shape: impressions rising while clicks flatten on queries where an AI Overview triggers. It is a signal, not a measurement, and it should be labelled that way in any report. For the Gemini app there is no equivalent telemetry at all — sampling prompts and recording what gets cited is the only honest method, which is what LLM visibility tools exist to automate.

Gemini SEO versus ChatGPT SEO

The two disciplines share a goal and almost nothing else. ChatGPT's citation path runs through crawlers you can allow or block independently of each other, and OpenAI documents them separately — so the technical decision is explicit and auditable. Gemini's runs through the Google index you already have, with a single opt-out that affects only the app. Practically: Gemini SEO has fewer switches and less to get wrong, but also far less to control. Most of the leverage is in the content, not the configuration. The ChatGPT-specific half is covered in ChatGPT SEO · Copilot SEO.

Questions we hear

How to rank in Gemini

Be indexed, stay eligible for a snippet, and write self-contained passages that answer specific questions. For the Gemini app specifically, do not disallow Google-Extended in robots.txt, since that token controls whether your pages can be used to ground the app's answers. There is no submission, no separate index, and no documented Gemini-only ranking factor.

Does blocking Google-Extended remove me from AI Overviews?

No. Google documents that Google-Extended “does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search.” AI Overviews are a Search feature governed by Googlebot access and the snippet controls. Blocking the token affects grounding in Gemini Apps, not your eligibility in Search.

Is Gemini SEO different from SEO?

For the Search surfaces, no — Google states there are no additional requirements or special optimizations needed for AI Overviews or AI Mode. The one genuinely separate decision is the Google-Extended token, which governs the Gemini app rather than Search. Everything else is ordinary technical and content quality work.

Does Gemini use the Google Search index?

For grounding, yes. Google describes grounding as “providing content from the Google Search index to the model at prompt time to improve factuality and relevancy” in Gemini Apps. That is why being indexed matters for the app as well as for Search, and why a page blocked from Googlebot is invisible to both.

Do I need an llms.txt file or AI-specific schema?

Not for Google's AI features. The documentation is explicit: “You don't need to create new machine readable files, AI text files, or markup to appear in these features. There's also no special schema.org structured data that you need to add.” Structured data still earns its place for rich results — it is simply not the entry ticket here.

Why do AI Overviews not appear for my keywords?

Often because Google chose not to show one. The documentation says AI Overviews “are only shown when our systems determine that it is additive to classic Search, and as such, often don't trigger.” Absence on a query is a signal about the query, not a defect on your page.

Go deeper

Related: generative engine optimization · what GEO means in SEO · ChatGPT SEO · LLM visibility tools · augmented intelligence SEO.

Want the eligibility check, the passage rewrite and the sampling run done for you? See how AI search visibility works as a module, or request a free AI-powered SEO audit as a first read.